JP5498109B2 - Defect detection apparatus and defect detection method - Google Patents

Description

  The present invention relates to a defect detection apparatus and a defect detection method.

  2. Description of the Related Art Conventionally, there has been provided an appearance inspection apparatus that inspects the presence or absence of defects on the appearance of an inspection object by performing binarization processing and differentiation processing on a captured inspection image (see, for example, Patent Document 1). In such an appearance inspection apparatus, one threshold value is set for the entire image, and it is possible to determine whether or not the inspection object is defective by binarizing the gray value of each pixel based on this threshold value. It has become.

  In addition, there is also provided an inspection image that is subjected to floating binarization processing to inspect for the presence or absence of defects. In this appearance inspection apparatus, the mode value of the gray value of the pixel is calculated for each predetermined line of the inspection image, and a threshold value is set for each line based on each calculated mode value. Then, it is possible to determine whether or not the inspection object is defective by binarizing the grayscale values of the pixels included in each line on the basis of the threshold value set for each line.

Japanese Patent No. 3890844 (paragraph [0020] -paragraph [0036] and FIGS. 1-15)

  For example, in the case of inspecting the presence or absence of defects in the molded part with respect to an object to be inspected such as a connector constituted by a molded part and a metal part held by the molded part, the above-mentioned former visual inspection apparatus has one entire image. Since the threshold value is set, for example, there is a case where a reflection unevenness generated in a curved surface portion of a metal part is erroneously detected as a defect in the formed part.

  In addition, since the threshold value is set for each line in the latter visual inspection apparatus described above, a defect in the molded part can be detected, but a part having a low contrast due to metal noise or the like is erroneously detected as a defect in the molded part. It was something to end up with.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a defect detection apparatus and a defect detection method that improve the detection accuracy of defects in a molded part.

The invention according to claim 1 is a defect detection apparatus for detecting a defect of an inspection object composed of a forming part and a metal part held by the forming part, and a predetermined inspection of the inspection object An imaging unit that captures an area; and a mode value of a gray value of a pixel for each predetermined line with respect to a partial image corresponding to the metal part of an inspection image captured by the imaging unit, and each calculated After setting a threshold value for each line from the mode value, a defect candidate pixel is extracted for each line by binarization based on the set threshold value, and a labeling process is performed on the extracted defect candidate pixel. Is executed to set a defect candidate portion, and an image processing unit that calculates an average gray value of the set defect candidate portion, and a preset range for determining whether or not the defect candidate portion is a defect are stored. a storage unit that, before Symbol defect candidate part is missing And a defect determination unit for determining whether the image processing unit calculates the average gray value of the corresponding partial image to the forming portion of the test image, the defect determination unit, the forming unit when the difference of the average gray value of said average gray value of the corresponding partial image and the defect candidate part is within the range of the preset, wherein the defect candidate part is determined to be defective .

The invention of claim 2, wherein the defect determination unit, when the defect candidate part is adjacent to the forming unit, the defect candidate parts is characterized in that it is determined that the defect.

The invention according to claim 3, wherein the image processing unit calculates a variation of the gray value of the defect candidate parts, the defect determination unit, the variation of the gray value of the defect candidate part below a preset threshold value in some cases, the defect candidate parts is characterized in that it is determined that the defect.

The invention according to claim 4, wherein the image processing unit calculates a variation of the gray value of the corresponding partial image to the forming portion of the test image, the defect determination unit, the variation of the gray value of the partial image If within the scope of the difference of the variation of the gray value is set in advance of the defect candidate parts, the defect candidate parts is characterized in that it is determined that the defect.

The invention according to claim 5 is a defect detection method for detecting a defect of an inspection object composed of a forming part and a metal part held by the forming part, and a predetermined inspection of the inspection object An inspection area imaging step of imaging an area by an imaging unit, and a mode value of a gray value of a pixel for each predetermined line with respect to a partial image corresponding to the metal part in an inspection image captured by the imaging unit A mode value calculating step, a threshold value is set for each line from each calculated mode value, and defect candidate pixels are extracted for each line by binarization based on the set threshold value an extraction step, the extracted running labeling processing on the defective pixel candidate setting the defect candidate part, and the average gray value calculation step of calculating an average gray value of the defect candidate part that is set before Symbol defect candidate Defects that determine whether a part is defective And a cross-sectional step, the calculated at the average gray value calculation step the average gray value of the corresponding partial image to the forming portion of the test image, said of the partial image corresponding to the molding portion in the defect decision step When the difference between the average gray value and the average gray value of the defect candidate portion is within a preset range, it is determined that the defect candidate portion is a defect.

  According to the invention of claim 1, by comparing the average gray value of the defect candidate portion with the defect threshold value, it becomes possible to discriminate between the defect of the formed portion and the metal noise of the metal portion, thereby reducing false detection. As a result, it is possible to improve the accuracy of detecting defects in the molded part. Moreover, there is an effect that productivity can be improved by reducing false detection.

In addition , by comparing the average gray value of the defect candidate part with the average gray value of the molding part, it becomes possible to make a comparison in accordance with the actual product, so that false detection can be further reduced, and as a result, detection accuracy is further improved. There is an effect that can be made.

According to the invention of claim 2 , when the defect candidate part is adjacent to the molding part, the defect candidate part is highly likely to be a defect, and the comparison result of the average gray value of the defect candidate part and the molding part is By evaluating together, it is possible to reduce erroneous detection as compared with the first aspect, and as a result, there is an effect that detection accuracy can be further improved.

According to the invention of claim 3 , when the defect candidate portion is caused by metal noise, the variation of the gray value becomes large. Therefore, by comparing the variation of the gray value with the predetermined threshold, the defect candidate portion It is possible to determine whether or not it is due to metal noise, and as a result, the detection accuracy can be further improved.

According to the invention of claim 4, by comparing the variation of the gray value of the defect candidate parts and the variation of the gray value of the molded part, erroneously compared to claim 3 since it becomes possible to compare in line with actual detection As a result, the detection accuracy can be further improved.

According to the invention of claim 5 , there is an effect that it is possible to realize a defect detection method with high detection accuracy capable of discriminating between defects in the formed part and metal noise in the metal part. In addition, by comparing the average gray value of the defect candidate part with the average gray value of the molding part, it becomes possible to make a comparison in accordance with the actual product, so that false detection can be further reduced, and as a result, detection accuracy is further improved. There is also an effect that can be made.

It is a schematic block diagram which shows the defect inspection apparatus of Embodiment 1. It is a cross-sectional perspective view of the connector which is an example of a test object same as the above. (A) is an inspection image same as the above, and (b) and (c) are data diagrams showing the gray value of a partial image corresponding to a metal part in the inspection image. It is a flowchart same as the above. It is an inspection image of the defect inspection apparatus of Embodiment 2. It is a flowchart same as the above. It is a flowchart of the defect inspection apparatus of Embodiment 3. It is a flowchart of the defect inspection apparatus of Embodiment 4. It is a flowchart of the defect inspection apparatus of Embodiment 5.

  Hereinafter, embodiments of a defect detection apparatus and a defect detection method according to the present invention will be described with reference to the drawings. The defect detection apparatus according to the present invention is used to detect a manufacturing defect of an inspection object (for example, a connector described later).

(Embodiment 1)
FIG. 1 is a schematic block diagram illustrating a defect detection apparatus A according to the first embodiment. The defect detection apparatus A includes, for example, an imaging unit 1 that includes a CCD camera and images a predetermined inspection area of an inspection object B, and an imaging unit. 1, an image processing unit 2 that executes image processing to be described later on the inspection image P <b> 1 (see FIG. 3A), a storage unit 3 for storing a defect threshold to be described later, and the storage unit 3. And a defect determination unit 4 for determining whether or not the inspection object B is defective based on the stored defect threshold.

  FIG. 2 is a cross-sectional perspective view of a connector 5 which is an example of an inspection object B. The connector 5 is a connector for electrically and mechanically connecting a pair of printed circuit boards (not shown), for example. Thus, it is composed of a vertically long molded part 5a made of a synthetic resin molded product and a metal part 5b made of a plurality of metal terminals arranged at equal intervals along the longitudinal direction of the molded part 5a. Since the connector 5 is well known in the art, a detailed description thereof will be omitted. The connector 5 is a connector mounted on one printed circuit board, and illustration and description of the connector mounted on the other printed circuit board are omitted.

  FIG. 3A shows an inspection image P1 imaged by the imaging means 1, and this inspection image P1 is obtained by imaging a part of the connector 5 from the direction of arrow C in FIG. 3 (b) and 3 (c) are data diagrams showing the gray value of each pixel when a partial image corresponding to the metal part 5b in the inspection image P1 is set in the inspection area a1, Numerical values in the pixels indicate the respective shade values. The inspection area a1 may be configured to be automatically set around a predetermined reference point, for example, or may be configured to be manually set while looking at a monitor or the like.

  By the way, in the image processing unit 2 described above, as shown in FIG. 3B, the mode value of the gray value of the pixel for each line af with respect to the partial image corresponding to the metal portion 5b (FIG. 3B). ), And a threshold value for binarization is set for each of the lines a to f from the calculated mode values. Further, the gray value of each pixel is binarized on the basis of each threshold value, and defect candidate pixels (shaded portions in FIG. 3B) having a gray value lower than the threshold value are extracted for each of the lines a to f. Finally, a labeling process is performed on the extracted defect candidate pixels to set defect candidate portions b1 and b2 (see FIG. 3C), and an average gray value of each defect candidate portion b1 and b2 is calculated. . In the present embodiment, the average gray value of the defect candidate part b1 is (80 + 81 + 79 + 80 + 81 + 79 + 80 + 80) ÷ 8 = 80, and the average gray value of the defect candidate part b2 is (105 + 100 + 90 + 120) ÷ 4≈103. In the present embodiment, four neighboring pixels are set in the same defect candidate portion for each extracted defect candidate pixel, and each defect candidate portion is labeled. For example, in the present embodiment, the label b1 is assigned to the upper left defect candidate portion in FIG. 3B, and the label b2 is assigned to the lower right defect candidate portion. Here, the vicinity of 4 means a pixel adjacent to a certain pixel in the vertical and horizontal directions.

  Further, a defect threshold value (for example, set to 100 as a value for discriminating forming burr and metal noise in this embodiment) is stored in the storage unit 3 in advance to determine whether or not the defect candidate portion is a defect. The defect determination unit 4 determines whether or not the defect candidate portion is defective by comparing the average gray value of the defect candidate portion with the defect threshold value. For example, in this embodiment, since the average gray value of the defect candidate portion b1 is 80 and is smaller than the defect threshold 100, the defect candidate portion b1 is determined to be a molding burr. On the other hand, the average gray value of the defect candidate portion b2 is 103, which is larger than the defect threshold 100, so that the defect candidate portion b2 is determined not to be a molding burr. Here, the defect candidate part b1 is a molding burr 5c (see FIG. 3A) of the molding part 5a, and the defect candidate part b2 is a metal noise (for example, uneven reflection) of the metal part 5b. According to the defect detection apparatus A of the embodiment, it becomes possible to discriminate between a defect (molding burr 5c) of the molded part 5a and a metal noise of the metal part 5b, so that false detection can be reduced, and as a result, Defect detection accuracy can be improved. In addition, productivity can be improved by reducing false detection.

  Next, the operation of the defect detection apparatus A will be described with reference to the flowchart of FIG. First, an inspection image P1 of the connector 5 is imaged by the imaging means 1 (step S1), and a partial image corresponding to the metal part 5b is set in the inspection area a1 in the captured inspection image P1 (step S2). Next, the image processing unit 2 calculates the mode value of the gray value of the pixel for each line a to f with respect to the partial image (step S3), and a predetermined offset value for each mode value. And a threshold value is set for each of the lines a to f. Furthermore, the image processing unit 2 binarizes the gray value of each pixel based on each threshold value, extracts defect candidate pixels for each of the lines a to f (step S4), and then extracts the defect candidate pixels that have been extracted. Then, the labeling process is executed to set the defect candidate parts b1 and b2, and the average gray value of the defect candidate parts b1 and b2 is calculated (step S5). And the defect judgment part 4 compares the average gray value of each defect candidate part b1, b2 with the defect threshold value memorize | stored in the memory | storage part 3, and the average gray value of each defect candidate part b1, b2 is more than a defect threshold value. If it is low, it is determined that it is a molding burr (step S6). In the present embodiment, the defect determination unit 4 determines that the defect candidate portion b1 is a molding burr and excludes the defect candidate portion b2 as not being a molding burr, so that only the molding burr can be extracted as a defect.

  Here, the above-described step S1 is an inspection region imaging step, step S3 is a mode value calculation step, step S4 is a defect candidate pixel extraction step, step S5 is an average gray value calculation step, and step S6 is a defect determination step. According to the embodiment, it is possible to realize a defect detection method with high detection accuracy capable of discriminating a defect (molding burr 5c) of the molded part 5a and metal noise of the metal part 5b.

  Note that the above defect threshold is an example, and may be set as appropriate according to the inspection object.

(Embodiment 2)
Embodiment 2 of the defect detection apparatus according to the present invention will be described. In the first embodiment, it is determined whether or not the defect candidate parts b1 and b2 are defective by comparing the average gray value of the defect candidate parts b1 and b2 with a defect threshold value stored in the storage unit 3 in advance. In the present embodiment, the average gray value of the partial image corresponding to the molding part 5a in the inspection image P1 is calculated, and the average gray value of the defect candidate parts b1 and b2 is compared with the average gray value of the molding part 5a. It is determined whether or not the defect candidate parts b1 and b2 are defective. In addition, about another structure, it is the same as that of Embodiment 1, the same code | symbol is attached | subjected to the same component and description is abbreviate | omitted. The following description will be given with reference to FIGS.

  The defect detection apparatus A of the present embodiment includes an imaging unit 1, an image processing unit 2, a storage unit 3, and a defect determination unit 4.

  The image processing unit 2 calculates the average gray value of the defect candidate portions b1 and b2 (80 and 103, respectively from the first embodiment) through the same processing as in the first embodiment, and further in the molding unit 5a in the inspection image P1. The corresponding partial image is set in the inspection area a2 (see FIG. 5), and the average gray value of the inspection area a2 is calculated. In the following description, it is assumed that the average gray value of the inspection area a2 is 90. Then, the defect determination unit 4 takes the difference between the average gray value of the defect candidate parts b1 and b2 and the average gray value of the inspection area a2, and if this difference is within a preset range (for example, ± 10). It is determined that the burr is a molding burr, and only the molding burr is extracted as a defect. Note that the allowable range of the difference is stored in the storage unit 3 in advance.

  Next, the operation of the defect detection apparatus A will be described with reference to the flowchart of FIG. First, an inspection image P1 of the connector 5 is imaged by the imaging means 1 (step S1), and a partial image corresponding to the metal part 5b is set in the inspection area a1 in the captured inspection image P1 (step S2). Next, the image processing unit 2 calculates the mode value of the gray value of the pixel for each line a to f with respect to the partial image (step S3), and a predetermined offset value for each mode value. And a threshold value is set for each of the lines a to f. Furthermore, the image processing unit 2 binarizes the gray value of each pixel based on each threshold value, extracts defect candidate pixels for each of the lines a to f (step S4), and then extracts the defect candidate pixels that have been extracted. Then, the labeling process is executed to set the defect candidate parts b1 and b2, and the average gray value of the defect candidate parts b1 and b2 is calculated (step S5). The image processing unit 2 sets a partial image corresponding to the forming unit 5a in the inspection image P1 in the inspection area a2 (step S6), and calculates an average gray value of the inspection area a2 (step S7). Then, the defect judgment unit 4 takes the difference between the average gray value of the defect candidate portions b1 and b2 and the average gray value of the inspection area a2, and if this difference is within a predetermined range (± 10), the defect judgment unit 4 It is determined that there is (step S8). In the present embodiment, since the difference in the average gray value between the defect candidate part b1 and the inspection area a2 is 10, it is determined as a molding burr, and the difference in the average gray value between the defect candidate part b2 and the inspection area a2 is 13. Therefore, only molding burrs can be extracted as defects.

  Thus, according to the present embodiment, by comparing the average gray value of the defect candidate portions b1 and b2 with the average gray value of the inspection area a2 corresponding to the forming portion 5a, comparison according to the actual product becomes possible. Therefore, erroneous detection can be reduced compared to the first embodiment, and as a result, detection accuracy can be further improved.

  The range of the difference is an example, and may be set as appropriate according to the inspection object.

(Embodiment 3)
Embodiment 3 of the defect detection apparatus according to the present invention will be described. In the second embodiment, it is determined whether or not the defect candidate parts b1 and b2 are defective by comparing the average gray value of the defect candidate parts b1 and b2 with the average gray value of the inspection area a2 corresponding to the molding part 5a. However, in this embodiment, the position information of the defect candidate parts b1 and b2 is further calculated, and it is determined whether or not the defect is based on the position information and the comparison result. In addition, about another structure, it is the same as that of Embodiment 2, and attaches | subjects the same code | symbol to the same component, and abbreviate | omits description. Further, the following description will be given with reference to FIGS. 1 to 3 and FIG.

  The defect detection apparatus A of the present embodiment includes an imaging unit 1, an image processing unit 2, a storage unit 3, and a defect determination unit 4.

  The image processing unit 2 calculates the average gray value of the defect candidate portions b1 and b2 and the average gray value of the inspection area a2 corresponding to the forming unit 5a through the same processing as in the second embodiment, and further each defect candidate portion. The position information of b1 and b2 is calculated. In the defect determination unit 4, the difference between the average gray values of the defect candidate portions b1 and b2 and the inspection area a2 is within a predetermined range (for example, ± 10), and the defect candidate portions b1 and b2 are formed in the molding unit 5a. If it is in contact with the metal, it is judged as a molding burr, and only the molding burr is extracted as a defect. The allowable range of the difference is stored in advance in the storage unit 3 as in the second embodiment. Here, the position information is the coordinates of the contour portions of the defect candidate portions b1 and b2, and when the coordinates of the contour portions are close to the contour coordinates of the forming portion 5a, it is determined that it is a forming burr.

  Next, the operation of the defect detection apparatus A will be described with reference to the flowchart of FIG. First, an inspection image P1 of the connector 5 is imaged by the imaging means 1 (step S1), and a partial image corresponding to the metal part 5b is set in the inspection area a1 in the captured inspection image P1 (step S2). Next, the image processing unit 2 calculates the mode value of the gray value of the pixel for each line a to f with respect to the partial image (step S3), and a predetermined offset value for each mode value. And a threshold value is set for each of the lines a to f. Furthermore, the image processing unit 2 binarizes the gray value of each pixel based on each threshold value, extracts defect candidate pixels for each of the lines a to f (step S4), and then extracts the defect candidate pixels that have been extracted. Then, the labeling process is executed to set the defect candidate parts b1 and b2, and the average gray value of each defect candidate part b1 and b2 is calculated and the position information is calculated (step S5). The image processing unit 2 sets a partial image corresponding to the forming unit 5a in the inspection image P1 in the inspection area a2 (step S6), and calculates an average gray value of the inspection area a2 (step S7). Then, the defect determination unit 4 takes the difference between the average gray value of the defect candidate portions b1 and b2 and the average gray value of the inspection area a2, and the difference is within a predetermined range (± 10), and each defect When the candidate parts b1 and b2 are in contact with the molding part 5a, it is determined that it is a molding burr (step S8). In the present embodiment, the difference in average gray value between the defect candidate part b1 and the inspection area a2 is 10, and the defect candidate part b1 is in contact with the molding part 5a, so that it is determined as a molding burr. On the other hand, the difference in the average gray value between the defect candidate part b2 and the inspection area a2 is 13, and since the defect candidate part b2 is not in contact with the molding part 5a, it is excluded as not being a molding burr. As a result, only molding burrs can be extracted as defects.

  Thus, according to the present embodiment, when the defect candidate part is adjacent to the molding part, the defect candidate part is likely to be a defect, and the average gray value of the defect candidate part and the molding part is compared. By evaluating together with the result, it is possible to reduce false detection as compared with the second embodiment, and as a result, detection accuracy can be further improved.

(Embodiment 4)
Embodiment 4 of the defect detection apparatus according to the present invention will be described. In the third embodiment, whether or not the defect is based on the comparison result of the average gray value of the defect candidate parts b1 and b2 and the average gray value of the inspection area a2 corresponding to the forming part 5a and the positional information of each defect candidate part b1 and b2. In this embodiment, the variation of the gray value of the defect candidate parts b1 and b2 (hereinafter referred to as standard deviation) is further calculated, and based on the standard deviation, the comparison result, and the position information. It is determined whether or not it is a defect. In addition, about another structure, it is the same as that of Embodiment 3, and attaches | subjects the same code | symbol to the same component, and abbreviate | omits description. Further, the following description will be given with reference to FIGS. 1 to 3 and FIG.

  The defect detection apparatus A of the present embodiment includes an imaging unit 1, an image processing unit 2, a storage unit 3, and a defect determination unit 4.

The image processing unit 2 calculates the average gray value, position information of the defect candidate portions b1 and b2, and the average gray value of the inspection area a2 corresponding to the forming unit 5a through the same processing as in the third embodiment. The standard deviation of the defect candidate parts b1 and b2 is calculated. Then, the defect determination unit 4 takes the difference between the average gray value of the defect candidate portions b1 and b2 and the average gray value of the inspection area a2, and the difference is within a predetermined range, and the defect candidate portions b1 and b2 Is in contact with the forming portion 5a and the standard deviation is equal to or less than a preset threshold value (for example, set to 5 in the present embodiment), it is determined as a forming burr, and only the forming burr is extracted as a defect. Here, the standard deviation σ is
From the equation (1), the standard deviation σ = 0.76 of the defect candidate part b1 and the standard deviation σ = 12.5 of the defect candidate part b2 are obtained. The above threshold value is stored in the storage unit 3 in advance.

  Next, the operation of the defect detection apparatus A will be described with reference to the flowchart of FIG. First, an inspection image P1 of the connector 5 is imaged by the imaging means 1 (step S1), and a partial image corresponding to the metal part 5b is set in the inspection area a1 in the captured inspection image P1 (step S2). Next, the image processing unit 2 calculates the mode value of the gray value of the pixel for each line a to f with respect to the partial image (step S3), and a predetermined offset value for each mode value. And a threshold value is set for each of the lines a to f. Furthermore, the image processing unit 2 binarizes the gray value of each pixel based on each threshold value, extracts defect candidate pixels for each of the lines a to f (step S4), and then extracts the defect candidate pixels that have been extracted. The labeling process is executed to set defect candidate parts b1 and b2, and the average gray value and position information of each defect candidate part b1 and b2 are calculated and the standard deviation is calculated (step S5). The image processing unit 2 sets a partial image corresponding to the forming unit 5a in the inspection image P1 in the inspection area a2 (step S6), and calculates an average gray value of the inspection area a2 (step S7). Then, the defect determination unit 4 takes a difference between the average gray value of the defect candidate portions b1 and b2 and the average gray value of the inspection area a2, and the difference is within a predetermined range (for example, ± 10), and each If the defect candidate parts b1 and b2 are in contact with the forming part 5a and the standard deviation is equal to or less than the preset threshold value 5, it is determined that the defect is a forming burr (step S8). In the present embodiment, the difference in average gray value between the defect candidate part b1 and the inspection area a2 is 10, the defect candidate part b1 is in contact with the forming part 5a, and the standard deviation of the defect candidate part b1 is 0.76. Since (<5), it is determined that the defect candidate part b1 is a molding burr. On the other hand, the difference in average gray value between the defect candidate part b2 and the inspection area a2 is 13, the defect candidate part b2 is not in contact with the forming part 5a, and the standard deviation of the defect candidate part b2 is 12.5 (> 5), the defect candidate part b2 is excluded as not being a molding burr. As a result, only molding burrs can be extracted as defects.

  Thus, according to the present embodiment, when the defect candidate portion is caused by metal noise, the variation in gray value (standard deviation) is large, whereas in the case of molding burr, the variation in gray value is Therefore, it is possible to determine whether or not the defect candidate portion is caused by metal noise by comparing the variation of the gray value with a predetermined threshold value, and as a result, the detection accuracy can be further improved.

  The threshold value is an example, and may be set as appropriate according to the inspection object.

(Embodiment 5)
Embodiment 5 of the defect detection apparatus according to the present invention will be described. In the fourth embodiment, the comparison result of the average gray value of the defect candidate parts b1 and b2 and the average gray value of the inspection area a2 corresponding to the forming part 5a, the position information of each defect candidate part b1 and b2, and each defect candidate part Although it is determined whether or not the defect is based on the comparison result with the standard deviation threshold values of b1 and b2, in this embodiment, the standard deviation of the inspection area a2 is calculated, and this standard deviation and each defect candidate part b1, It is determined whether or not it is a defect based on the comparison result of the standard deviation of b2, the comparison result of the average gray value, and the position information. In addition, about another structure, it is the same as that of Embodiment 4, the same code | symbol is attached | subjected to the same component and description is abbreviate | omitted. Further, the following description will be given with reference to FIGS. 1 to 3 and FIG.

  The defect detection apparatus A of the present embodiment includes an imaging unit 1, an image processing unit 2, a storage unit 3, and a defect determination unit 4.

  The image processing unit 2 calculates the average gray value, position information, and standard deviation of the defect candidate portions b1 and b2 through the same processing as in the fourth embodiment, and also calculates the average gray value of the inspection area a2 corresponding to the forming unit 5a. And the standard deviation of the inspection area a2 is calculated. In the following description, it is assumed that the average gray value of the inspection area a2 is 90 and the standard deviation is 0.7. In the defect determination unit 4, the difference between the average gray value of the defect candidate portions b1 and b2 and the average gray value of the inspection area a2 is within a predetermined range (for example, ± 10), and the defect candidate portions b1 and b2 Is determined to be a molding burr when the difference between the standard deviation of each of the defect candidate parts b1 and b2 and the standard deviation of the inspection area a2 is within a predetermined range (for example, ± 2). Only molding burrs are extracted as defects. The standard deviation of each defect candidate part b1, b2 is obtained from the equation (1) as in the fourth embodiment. The standard deviation σ = 0.76 of the defect candidate part b1 and the standard deviation σ = 12 of the defect candidate part b2 .5. Further, the allowable range of the difference is stored in the storage unit 3 in advance.

  Next, the operation of the defect detection apparatus A will be described with reference to the flowchart of FIG. First, an inspection image P1 of the connector 5 is imaged by the imaging means 1 (step S1), and a partial image corresponding to the metal part 5b is set in the inspection area a1 in the captured inspection image P1 (step S2). Next, the image processing unit 2 calculates the mode value of the gray value of the pixel for each line a to f with respect to the partial image (step S3), and a predetermined offset value for each mode value. And a threshold value is set for each of the lines a to f. Furthermore, the image processing unit 2 binarizes the gray value of each pixel based on each threshold value, extracts defect candidate pixels for each of the lines a to f (step S4), and then extracts the defect candidate pixels that have been extracted. The labeling process is executed to set the defect candidate parts b1 and b2, and the average gray value, position information, and standard deviation of the defect candidate parts b1 and b2 are calculated (step S5). Further, the image processing unit 2 sets a partial image corresponding to the molding unit 5a in the inspection image P1 in the inspection area a2 (step S6), and calculates an average gray value and a standard deviation of the inspection area a2 (step S7). ). Then, the defect determination unit 4 takes the difference between the average gray value of the defect candidate portions b1 and b2 and the average gray value of the inspection area a2, and the difference is within a predetermined range (± 10), and each defect If the candidate parts b1 and b2 are in contact with the molding part 5a and the difference between the standard deviation of each defect candidate part b1 and b2 and the standard deviation of the inspection area a2 is within a predetermined range (± 2), molding burr It is determined that there is (step S8). In the present embodiment, the difference in average gray value between the defect candidate part b1 and the inspection area a2 is 10, the defect candidate part b1 is in contact with the forming part 5a, and the standard deviation between the defect candidate part b1 and the inspection area a2 Is 0.06 (<2), the defect candidate part b1 is determined to be a molding burr. On the other hand, the difference in the average gray value between the defect candidate part b2 and the inspection area a2 is 13, the defect candidate part b2 is not in contact with the forming part 5a, and the difference in standard deviation between the defect candidate part b2 and the inspection area a2 Since 11.8 (> 2), the defect candidate part b2 is excluded as not being a forming beam. As a result, only molding burrs can be extracted as defects.

  Thus, according to the present embodiment, the variation (standard deviation) of the grayscale values of the defect candidate portions b1 and b2 is compared with the variation of the grayscale values of the inspection area a2 corresponding to the molding portion 5a, so that Therefore, the erroneous detection can be reduced as compared with the fourth embodiment, and as a result, the detection accuracy can be further improved.

  The range of the difference is an example, and may be set as appropriate according to the inspection object.

  In the first to fifth embodiments described above, the case where the inspection object B is a connector has been described as an example. However, the inspection object is not limited to a connector, for example, other than semiconductor components such as an integrated circuit and a sensor. Electronic components such as relays and sockets may also be used. Furthermore, in the first to fifth embodiments described above, the metal part 5b has been described as an example. However, the metal part is not limited to the metal part as long as it is glossy, and other parts may be used.

DESCRIPTION OF SYMBOLS 1 Imaging means 2 Image processing part 3 Memory | storage part 4 Defect judgment part 5 Connector 5a Molding part 5b Metal part A Defect detection apparatus B Inspected object

Claims (5)

A defect detection device for detecting a defect of an object to be inspected composed of a molded part and a metal part held by the molded part,
An imaging unit that images a predetermined inspection area of the inspection object, and a mode of a gray value of a pixel for each predetermined line with respect to a partial image corresponding to the metal portion of the inspection image captured by the imaging unit After calculating a value and setting a threshold value for each line from each calculated mode value, the defect candidate pixels are extracted for each line by binarization based on the set threshold value, and the extracted In order to determine whether or not the defect candidate portion is defective, and an image processing portion that performs a labeling process on the defect candidate pixel to set a defect candidate portion, calculates an average gray value of the set defect candidate portion, and comprising of a storage unit for storing a preset range, before Symbol defect candidate part and a defect determination unit for determining whether or not a defect,
The image processing unit calculates an average gray value of a partial image corresponding to the molding unit in the inspection image, and the defect determination unit is configured to calculate the average gray value of the partial image corresponding to the molding unit and the defect. when the difference of the average gray value of the candidate section is within the range of the preset, the defect detecting device, wherein the defect candidate part is determined to be defective.   The defect detection apparatus according to claim 1, wherein the defect determination unit determines that the defect candidate part is a defect when the defect candidate part is adjacent to the forming part.   The image processing unit calculates a variation in gray value of the defect candidate portion, and the defect determination unit determines that the defect candidate portion is less than or equal to a preset threshold value when the variation in gray value of the defect candidate portion is equal to or less than a preset threshold value. The defect detection apparatus according to claim 2, wherein it is determined that the defect is a defect.   The image processing unit calculates a variation in the gray value of the partial image corresponding to the forming unit in the inspection image, and the defect determination unit determines the variation in the gray value of the partial image and the defect candidate portion. The defect detection apparatus according to claim 3, wherein the defect candidate portion is determined to be a defect when a difference in gradation value variation is within a preset range. A defect detection method for detecting a defect of an object to be inspected composed of a molded part and a metal part held by the molded part,
An inspection area imaging step of imaging a predetermined inspection area of the object to be inspected by an imaging means, and a pixel for each predetermined line with respect to a partial image corresponding to the metal part of the inspection image captured by the imaging means A mode value calculation step for calculating the mode value of the gray value, a threshold value is set for each line from each calculated mode value, and binarization is performed on the basis of the set threshold value. A defect candidate pixel extraction step for extracting line by line, and an average gray value for calculating the average gray value of the set defect candidate part by performing a labeling process on the extracted defect candidate pixel to set a defect candidate part comprising a calculation step, before Symbol defect candidate part and a defect judgment step of judging whether or not a defect,
The average gray value of the partial image corresponding to the molded part in the inspection image is calculated in the average gray value calculating step, and the average gray value and the defect of the partial image corresponding to the molded part in the defect determining step. A defect detection method comprising: determining that the defect candidate portion is a defect when a difference in the average gray value of the candidate portion is within a preset range.